How do you get the evaluation results of the `prompt enginering` leaderboard?
zhimin-z opened this issue · 3 comments
Check https://promptbench.readthedocs.io/en/latest/leaderboard/pe.html:
I failed to locate any related report or paper that could reproduce these evaluation results...
@madhavMathur @jindongwang @msftgits @dnfclas
In the latest version of the paper, only Figure 4 is directly related to the leaderboard above:
However, this figure is just a subset of these evaluation results and there are no numeric values shown on the top of the chart bars. Would you mind adding more explanation to the benchmark documentation?
@icecream-and-tea can answer this question
For the sake of convenience in presentation, we only included the main part of the prompt engineering evaluation results in the paper. The evaluation of all data can be found in this page ( https://promptbench.readthedocs.io/en/latest/leaderboard/pe.html)
For the sake of convenience in presentation, we only included the main part of the prompt engineering evaluation results in the paper. The evaluation of all data can be found in this page ( https://promptbench.readthedocs.io/en/latest/leaderboard/pe.html)
Add a PR accordingly: #58